Prior to the introduction of Milan criteria, the long-term results of liver transplantation in patients with hepatocellular carcinoma have been variable and disappointing, with an overall 5-year survival rate ranging from 30 to 40 % as shown in Table 11.2 [9–16]. However, these studies showed a positive correlation between the tumor burden prior to transplant and post-transplant recurrence rates. In 1996, Mazzaferro et al. reported the outcomes of 48 patients, with cirrhosis and small, unresectable hepatocellular carcinomas , who underwent liver transplantation at the National Cancer Institute in Milan, Italy [5]. This study showed 4-year actuarial survival rates of 73 % and the recurrence-free survival rate of 83 %, and these observations were further corroborated by many other investigators (Table 11.3) [17–25]. These observations were further supported by the outcome data (“real-life data”) from the United Network for Organ Sharing (UNOS) database [26]. An analysis of UNOS data suggested that the survival did improve after the publication of Milan criteria in the United States [21]. Based on published literature, 5-year survival of patients transplanted for HCC based on Milan criteria is around 65–78 % whereas it is 68–75 % for non-tumor patients transplanted during the same period [27]. When HCC patients were transplanted outside Milan criteria, 5-year survival rates were 46–60 %. A detailed analysis of studies suggested the hazard ratio is 1.68 when transplanted outside Milan criteria, but hazard ratio was lower (1.28) for live donor liver transplant recipients [27].

Although Milan criteria are well validated, the cutoff for the size and the number of tumor nodules are rather arbitrary, and are decided by imaging studies, but not based on explant pathology or tumor biology. It has been suggested that Milan criteria are rather stringent, and moreover, imaging studies may underestimate or overestimate the tumor size and tumor numbers depending on pre-transplant radiological imaging techniques, imaging protocols, and more importantly on the interpretation skills of the radiologists. A retrospective analysis of the UNOS/OPTN database confirmed these concerns and showed that radiologic examinations were not very precise, when compared to explant pathology, underestimating tumor load in 27 % and overestimating in 30 % of the population [28]. It has been also suggested that the Milan criteria, based on pre-transplant radiologic criteria, were proposed more than two decades ago, and since then there have been significant improvements in both CT scan and MRI imaging techniques. Nevertheless, a recent review of literature suggested that Milan criteria are very robust in predicting excellent survival, and moreover, those who were transplanted within Milan criteria had more favorable tumor differentiation [27]. When interpreting the excellent outcomes of LT based on Milan criteria, it is important to remember that these results were not reported as “intention to treat” analysis, but based on those who received LT in a timely manner. During the waiting period, tumors do progress and many people may fall outside the Milan criteria and hence do not receive a liver transplantation.

Increasing evidence suggests that tumor biology, and not the number and size of tumor nodules, is the most important predictor of favorable outcome. However, until we identify reliable markers of unfavorable tumor biology, we have to depend on other surrogate markers for patient selection. Several risk factors, including significantly elevated alpha fetoprotein (AFP) levels , poor tumor differentiation, and vascular invasion, have been identified that could predict recurrence of HCC even when transplanted within Milan criteria. A higher AFP prior to LT is considered an important predictor of post-LT HCC recurrence in many studies. Higher AFP levels are often associated with more advanced cirrhosis, vascular invasion, higher tumor burden, and poor performance status. An analysis of UNOS database from region 5, where there is a longer waiting period before LT, showed that high AFP was the only pre-transplant variable that predicted post-transplant tumor recurrence and mortality in patients who underwent LT for HCC within Milan criteria [29]. Similarly, in a retrospective cohort study [30] of 313 HCC patients undergoing transplantation—pre-transplant AFP, lens culinaris agglutinin-reactive AFP (AFP-L3), and des-gamma-carboxy prothrombin (DCP) predicted HCC recurrence after transplantation. When compared to LT done within Milan criteria, hazard ratio (HR) were 2.6 (1.4–4.7, p = 0.003) for outside Milan, 8.6 (3.0–24.6, p < 0.0001) for outside Milan, and AFP ≥250 ng/mL and 7.2 (2.8–18.1, p < 0.0001) for outside Milan and DCP ≥7.5 ng/mL. These findings suggest that using both biomarkers and Milan criteria may be better than using the Milan criteria alone in determining liver transplantation eligibility.

Tumor differentiation and vascular invasion are other important predictors of HCC recurrence. In a study of 155 patients who underwent liver transplants for HCC (84 % within Milan criteria based on the explanted livers), histological grade of differentiation and macroscopic vascular invasion were strong independent predictors of survival [31]. Other investigators have also confirmed that histological differentiation and vascular invasion are independent predictors of survival [32, 33]. It is also important to note that poorly differentiated tumors are more likely to be associated with vascular invasion. Unlike AFP, histological grade of tumor differentiation is unknown before LT as it is not a common practice to do biopsy of HCC, that meet well-established diagnostic criteria of HCC, because of the potential risk for needle track metastases [34, 35]. Studies that have compared contrast-enhanced dynamic imaging of liver with the degree of histopathological differentiation of HCC have reported a good correlation between hypervascular enhancement patterns with higher pathological grades [36–40]. Other investigators, however, have reported a decline in arterial blood supply in the later stages of HCC progression (grade ≥3) [37–39]. A recent Italian retrospective study [41] showed that patients with hypovascular HCC have a lower tendency towards recurrence and a prolonged recurrence-free survival than those with hypervascular HCC. These studies suggest that dynamic imaging findings may not be reliable surrogate markers of tumor differentiation. Although microvascular invasion is a very strong and a consistent predictor for higher recurrence, currently there are no imaging techniques to diagnose microvascular invasion, and when found, it is almost always based on explant pathology.

There are few other predictors that could predict higher HCC recurrence. Recipients of grafts from older donors (≥60 years) or those who received organs through regional sharing have been reported to have significantly higher risk of HCC recurrence, but these data need further corroboration because of many confounding variables [42]. A retrospective Belgian study [43] suggested that FDG positron emission tomography computed tomography with a tumor/liver activity ratios (RSUV max) cutoff value of 1.15 or more is a strong prognostic factor for recurrence and death in patients with HCC treated by LT. In their study, none of the patients outside the MILAN criteria with RSUV max <1.15 suffered from recurrence in the follow-up. Chinese investigators have suggested that pre-transplant platelet to lymphocyte ratio (PLR) ≥125 or preoperative neutrophil-lymphocyte ratio (NLR) ≥4 could be associated with advanced tumor stage and behavior [44] and could be used as a predictor of post-transplant HCC recurrence [44, 45]. These observations need to be corroborated in larger, prospective studies after adjusting for other known confounders. In the future, we may have more refined molecular markers that could be used in association with imaging to better define the LT selection criteria for those with HCC.

Tumor recurrence is associated with a very poor prognosis, with median survival <12 months. Sites of recurrence include liver alone (16 %), both intra and extra hepatic (31 %), or extrahepatic alone (53 %). Since the introduction of Milan criteria, tumor recurrence rates have dropped from a median of 25.5 % down to 8–11 %. Although one could argue that aggressive surveillance of all patients after LT is probably not cost-effective, most centers do surveillance every 3 months in the first year, every 6 months in the second year, and every 6–12 months from years 3 to 5. Surveillance intervals could be tailored (less frequent if explant pathology is favorable) based on explant pathology to reduce the costs [46]. Although there is no consensus, authors prefer non-contrast CT of chest and contrast-enhanced MRI of abdomen for surveillance purposes. If patients had elevated AFP prior to LT, monitoring of AFP may be helpful. Other options include ultrasound and AFP every 3–6 months for 5 years, but the sensitivity of ultrasound may be less than optimal.

The level and type of immunosuppressive agents may play a role in tumor recurrence and progression after LT as shown in few experimental and clinical studies [47–49]. It has been suggested that a higher level of cyclosporine (CsA) exposure, especially during the first year after LT, may lead to higher tumor recurrence rates [50, 51]. Alternatively, mTOR inhibitors such as sirolimus may have an advantage over tacrolimus or cyclosporine in those who received LT for HCC because of the antiangiogenic properties of the drug [52–54]. mTOR is overexpressed by up to two-thirds in HCC, and in animal models, sirolimus has shown efficacy at reducing tumor growth and longer survival. Zimmerman et al. compared two groups of patients who underwent liver transplantation for HCC: patients on sirolimus and calcineurin inhibitor post-transplant vs. those on cyclosporine or tacrolimus plus mycophenolate mofetil and corticosteroids. The 1- and 5-year survival rates for the sirolimus-treated group were 95.5 % and 78 %, respectively, versus 83 % and 62 % for the non-sirolimus group [55]. In a retrospective case–control study [56], patients who received sirolimus and post-transplant chemotherapy had better recurrence-free survival than patients who were treated with tacrolimus and mycophenolate mofetil along with post-transplant chemotherapy. Although no firm conclusions can be based on these studies, it is perhaps prudent to offer sirolimus-based therapy to those who received LT for HCC.

The optimal treatment for local HCC recurrence is surgical resection, if possible. In one small study, involving 17 patients who had recurrent HCC, overall survival rates of the surgical group were similar to that of the patients without HCC recurrence [57]. If surgical resection is not feasible, as it in majority of cases, other ablative modalities such as radio frequency ablation or chemoembolization may be considered. The role of neoadjuvant and adjuvant treatment for recurrent HCC after transplantation has been discussed in detail in another chapter. Sorafenib is an oral multikinase inhibitor that blocks multiple growth factor pathways including vascular endothelial growth factor receptor (VEGFR) and platelet-derived growth factor receptor-β (PDGFR-β), and two registration trials (not in transplant recipients) showed that it can prolong life in those with unresectable HCC [58–60]. A retrospective study showed that sorafenib is tolerated by transplant recipients, with recurrent HCC, on sirolimus-based immunosuppressive regimen without any adverse effects on graft function [61]. In this study, common adverse events were diarrhea (46 %), hand-foot skin reaction (27 %), nausea, fatigue, and leucopoenia (18 %), and these adverse events are similar to those reported non-transplant patients. Another retrospective analysis on LT recipients with unresectable HCC recurrence and undergoing combination therapy with everolimus and sorafenib, adverse events led to drug discontinuation and dose reduction of sorafenib in two patients (28 %) and three (43 %), respectively [62]. Recently, a placebo-controlled, double-blind (STORM) trial [63] was designed to evaluate the efficacy and safety of adjuvant sorafenib in patients with HCC who have no lesions after curative resection or ablation. Treatment with sorafenib after curative resection or ablation of HCC did not improve recurrence-free survival when compared with placebo. Additionally, time to recurrence and overall survival showed no differences between the treatment arms. Discontinuation rates with sorafenib were higher (24 % vs. 7 %) compared to placebo. Based on this study, one may conclude that the “preventive” role of sorafenib is unproven.